The present invention relates to image sensors, and more particularly to pixels of image sensors.
Digital cameras are today one of the most popular consumer electronics. The “MegaPixel” capacity and color sensitivity of image sensors is currently a limiting factor limiting image quality of images captured by digital cameras. Image sensors function to transform (visible) light that is focused onto the image sensor by optical lenses, into electrical signals to capture and display images of high quality. Additionally, a very fast “shutter speed” or “film speed” equivalent in digital image sensors is a strong selling point for high-end digital cameras and can enhance image quality (e.g., reducing blurring due to subject or camera motion).
A typical image sensor comprises a pixel array that has a plurality (array) of pixels. Each pixel includes a photodiode for generating signal charges in response to photons (i.e., light) incident thereon, and an electronic element for transferring and outputting the signal charges from the photodiode. Depending upon the manner of transferring and outputting signals charges, image sensors are roughly classified into two kinds, i.e., charge-coupled devices (CCD; hereinafter, ‘CCD image sensors’) and complementary metal-oxide-semiconductor (CMOS) image sensor (hereinafter, ‘CMOS image sensor’). The CCD image sensor uses pluralities of MOS capacitors for accumulating, transferring and outputting charges. By applying appropriate voltages to the electrodes of the MOS capacitors, signal charges of each pixel are successively transferred by way of the MOS capacitors. The CMOS image sensor uses pluralities of transistors, by which signal charges generated by the photodiode are converted into a voltage at each pixel and output therefrom.
The CCD image sensors typically have better noise and image quality than CMOS image sensors, but CMOS image sensors typically have lower product cost and lower power consumption than CCD image sensors. In other words, the CMOS image sensor has the advantages of lower power, singularity of voltage and current, compatibility with combined CMOS circuits (e.g., integrated on the same chip), random access of image data, and lower production cost by employing the standard of CMOS technology. Thus, CMOS image sensors are gaining market share in various applications such as digital cameras, smart phones, personal digital assistants (PDA), notebook computers, security cameras, barcode detectors, high-definition television sets, children's toys, and so on.
FIG. 1 is a plane view of a portion of a pixel array of a conventional image sensor. Referring to FIG. 1, pixels are formed around active regions 13 of a semiconductor substrate, the active regions being electrically isolated from each other (like “semiconductor islands”), each of which includes a photodiode and pluralities of transistors. Each active region 13 may be sectored into a first active region 13A including the photodiode, and a second active region 13B including the plural transistors. On the second active region 13B are arranged a transfer gate 21, a reset gate 23, a source follower gate 25, and a selection gate 27. The transfer gate 21 is located adjacent to the first active region 13A. An impurity region formed in the second active region 13A between the transfer and reset gates 21 and 23 serves as a floating diffusion region 29 that is electrically connected to the source follower gate 25. An impurity region formed in the second active region 13B between the reset and source follower gates 23 and 25 acts as a reset diffusion region 31. There is an impurity region 33 in the second active region 13B between the source follower and selection gates 25 and 27, and an impurity region 35 in the second active region 13B outside of the selection gate 27. Each transistor is formed of a gate and the impurity regions at either side of its gate.
The vertically sectional structure of a representative pixel is next described with reference to FIG. 2. FIG. 2 is a cross-sectional view of one pixel taken along section line I-I′ of FIG. 1. Referring to FIG. 2, the photodiode includes an N-type region 17 and a P-type region 19 formed in the first active region 13A. The floating diffusion (FD) region 29 is electrically connected to the source follower gate 25 through a local interconnection 37.
In the general structure of a conventional pixel, it is necessary to form the photodiode and gates in the same active region 13 and to allocate a part (e.g., 13B) thereof to the transistors. Thus, only a portion, i.e., the first active region 13A of the active region 13 is used for the photodiode. Thus, there is a limitation of the conventional image sensor due to a fill factor that represents the area occupied in a pixel by the photodiode.
The “fill factor” indicates the size of the light sensitive photodiode relative to the entire pixel and is the fraction of the surface that is sensitive to light. A large fill factor is desirable because the larger the fill factor the more light will be captured by the chip up to the maximum of 100%. This helps improve the Signal-to-Noise Ration (SNR). Because of the extra electronics (e.g., transistors) required in each pixel the “fill factor” tends to be quite small, especially for Active Pixel Sensors which have more per pixel transistors. To overcome this limitation, some have proposed an array of microlenses to be placed on top of the sensors, hut this increases the production cost. Additionally, in the conventional image sensor, light incident upon a target pixel, especially, light incident from a slanted angle, may be reflected by an interconnection of the target pixel or a gate and may arrive at an adjacent pixel, causing across-talk affect or distortion.